838 
Journal oj Agricultural Research 
Vol. XXIV, No. lo 
of variation, indicates linkage in the two parents between the factors Yy 
for stipule width and the Aa factors for flower color. Recombinations 
in the F2 of cross-over and noncross-over gametes would account for 
differences in the range of variation and in the means between the 
F2 A and a segregates and the A and a parents. 
However, as the combined frequency distributions of the two crosses 
of Mummy with rogue cultures under discussion (0.1059 and 0.1093) 
show considerable variation from the F2 frequency distribution of 
Gradus as type X Mummy, especially in the position of the respective 
modes and means, it is unlikely that the factor allelomorphic to x' in 
the Fj hybrid is the x of Gradus type; it is apparently a mutation of X 
to some other factor for stipule shape, the mutation being analogous to 
the genetic change occurring when intergrading intermediates are pro¬ 
duced from Early Giant types (j). In this case the intermediates are 
of the sort producing few or no rogues among their progeny. 
Although the ranges of variation of the two atypical F2 progenies 
(0.1059 and 0.1093) overlaps the range of Gradus rogue, no rogues were 
present in the cultures judging from the character of the F3 generation. 
Six F2 plants of culture 0.1093 from classes 2.05, 2.15, and 2.45 of the 
variation curve gave only broad and intermediate F3 segregates. It 
would appear that the rogue factors which entered the Fj^ zygote failed 
to function as rogue factors at gametogenesis, having mutated to some¬ 
thing different from X. 
RELATION OF COLOR FACTORS A AND C TO STIPULE SHAPE IN THE TYPICAL Fj 
FAMILIES OF THE CROSSES OF MUMMY WITH GRADUS ROGUE 
The stipules of 1,404 plants, progeny of the remaining Fj families 
(omitting cultures 0.1059 were measured. The plants 
represented 49 families, of which the frequency distributions and the 
derived statistical constants are given in Table XV. 
Of the 1,404 individuals, 356 were white (of constitution aa); 305 pink 
(with A and B factors); and 873 were purple (with A, B, and C factors). 
The mean stipule ratio of the white segregates was found to be 
2.258±0.0110 as compared with 2.336±0.0110 for the pinks and 
2.325±0.0063 for tho purples (Table VII). 
As no essential difference exists between the means of the pink and 
purple F2 plants, and as they each contain the A factor, the two color 
populations may be combined for comparison with the aa segregates. 
The mean of the A A and Aa plants is 2.335 ±0.0056 contrasted with 
2.258 ±0.0110 for the aa segregates, a difference in the means of over 
6 times the probable error of the difference. 
For a comparison inter se of the frequency distributions of the AA and 
Aa, and aa segregates and of the F2 as a whole the number of variates 
in each class was expressed in percentages of the total number of variates 
in that particular category of plants. Figure 3 shows the curves obtained 
by plotting the percentage of variates in each class for the AA and Aa, 
and aa segregates separately, and for all the F2 plants irrespective of the 
color factors present. The graphs for Gradus, Gradus rogue, and Mummy 
are shown in figure 2, for comparison with the graphs of the F2 genera¬ 
tion. 
